This report explores a prosper loan data set containing 81 features and 113937 listings. The main interest is lender yield, which is the most important question to lenders: how much can I make if I invest in this loan? the other 10-15 features will be chosen based on intuition, common sense, a little experiment and the result of goggling and studying.
After July 2009, prosper implemented a new borrower rating system, and because this is a very important features in this analysis, so all the listing before July 2009 were removed. Which leaves us 84853 listings.
Summary of features that are picked to explore.
## [1] 84853 15
## 'data.frame': 84853 obs. of 15 variables:
## $ IncomeRange : Factor w/ 7 levels "Not employed",..: 5 4 7 7 4 4 4 4 6 4 ...
## $ ProsperRating..Alpha. : Factor w/ 8 levels "AA","A","B","C",..: 2 2 5 3 6 4 1 1 4 3 ...
## $ DebtToIncomeRatio : num 0.18 0.15 0.26 0.36 0.27 0.24 0.25 0.25 0.12 0.18 ...
## $ IsBorrowerHomeowner : Factor w/ 2 levels "FALSE","TRUE": 1 2 2 2 1 1 2 2 1 1 ...
## $ InquiriesLast6Months : int 3 0 1 0 0 3 1 1 1 1 ...
## $ TradesNeverDelinquent..percentage.: num 1 0.76 0.95 1 0.68 0.8 1 1 0.72 1 ...
## $ CreditScoreRangeLower : int 680 800 680 740 680 700 820 820 640 680 ...
## $ ProsperScore : num 7 9 4 10 2 4 9 11 7 4 ...
## $ OpenRevolvingMonthlyPayment : num 389 115 220 1410 214 101 219 219 25 290 ...
## $ BankcardUtilization : num 0.21 0.04 0.81 0.39 0.72 0.13 0.11 0.11 0.51 0.7 ...
## $ LoanCurrentDaysDelinquent : int 0 0 0 0 0 0 0 0 0 3 ...
## $ MonthlyLoanPayment : num 319 321 564 342 123 ...
## $ LP_InterestandFees : num 0 1052 1257 327 622 ...
## $ LenderYield : num 0.082 0.0874 0.1985 0.1214 0.2612 ...
## $ LoanOriginalAmount : int 10000 10000 15000 15000 3000 10000 10000 10000 13500 4000 ...
## IncomeRange ProsperRating..Alpha. DebtToIncomeRatio
## Not employed : 649 C :18345 Min. : 0.000
## $0 : 45 B :15581 1st Qu.: 0.150
## $1-24,999 : 4654 A :14551 Median : 0.220
## $25,000-49,999:24175 D :14274 Mean : 0.259
## $50,000-74,999:25627 E : 9795 3rd Qu.: 0.320
## $75,000-99,999:14498 HR : 6935 Max. :10.010
## $100,000+ :15205 (Other): 5372 NA's :7296
## IsBorrowerHomeowner InquiriesLast6Months
## FALSE:40005 Min. : 0.0000
## TRUE :44848 1st Qu.: 0.0000
## Median : 0.0000
## Mean : 0.9646
## 3rd Qu.: 1.0000
## Max. :27.0000
##
## TradesNeverDelinquent..percentage. CreditScoreRangeLower ProsperScore
## Min. :0.0800 Min. :600.0 Min. : 1.00
## 1st Qu.:0.8500 1st Qu.:660.0 1st Qu.: 4.00
## Median :0.9500 Median :700.0 Median : 6.00
## Mean :0.9059 Mean :699.4 Mean : 5.95
## 3rd Qu.:1.0000 3rd Qu.:720.0 3rd Qu.: 8.00
## Max. :1.0000 Max. :880.0 Max. :11.00
##
## OpenRevolvingMonthlyPayment BankcardUtilization LoanCurrentDaysDelinquent
## Min. : 0.0 Min. :0.0000 Min. : 0.00
## 1st Qu.: 156.0 1st Qu.:0.3300 1st Qu.: 0.00
## Median : 311.0 Median :0.6000 Median : 0.00
## Mean : 430.7 Mean :0.5642 Mean : 36.63
## 3rd Qu.: 564.0 3rd Qu.:0.8300 3rd Qu.: 0.00
## Max. :13765.0 Max. :2.5000 Max. :1593.00
##
## MonthlyLoanPayment LP_InterestandFees LenderYield LoanOriginalAmount
## Min. : 0.0 Min. : -2.35 Min. :0.0300 Min. : 1000
## 1st Qu.: 157.3 1st Qu.: 254.06 1st Qu.:0.1259 1st Qu.: 4000
## Median : 251.9 Median : 674.39 Median :0.1775 Median : 7500
## Mean : 291.9 Mean : 1042.20 Mean :0.1860 Mean : 9083
## 3rd Qu.: 388.4 3rd Qu.: 1436.06 3rd Qu.:0.2474 3rd Qu.:13500
## Max. :2251.5 Max. :10572.78 Max. :0.3400 Max. :35000
##
If you are a lender, then the LenderYield should be something you care the most. This number tells you how much you can make. However, if this number is accurate, higher LenderYield should represent higher risk. And judging from the histogram, there are large percentage of Prosper borrowers whose credit rating is not so great.
Distribution of the income range of the borrowers, 77520 out of 84853, or approximately 75%, their income is verifiable. Distribution of StatedMonthlyIncome is a typical right skewed distribution. I set the x-axis limit from 0 to 20000 because 99% of the borrowers state that their monthly income are less than 20000.
## 99% of stated incomes are less than 20000
## [1] 0.9888867
## 77520 stated incomes are verifiable
##
## FALSE TRUE
## 7333 77520
purpose of the borrowing, roughly 63% of the borrowing is for debt consolidation.
## percentage of each purpose of borrowing
##
## Not Available Debt Consolidation Home Improvement
## 0.0002357017 0.6267309347 0.0801503777
## Business Personal Loan Student Use
## 0.0624373917 0.0000000000 0.0032291139
## Auto Other Baby&Adoption
## 0.0263632400 0.1086349334 0.0023452323
## Boat Cosmetic Procedure Engagement Ring
## 0.0010017324 0.0010724429 0.0025573639
## Green Loans Household Expenses Large Purchases
## 0.0006953201 0.0235230340 0.0103237363
## Medical/Dental Motorcycle RV
## 0.0179369026 0.0035826665 0.0006128245
## Taxes Vacation Wedding Loans
## 0.0104298021 0.0090509469 0.0090863022
Debt to income ratio, most (99%) borrowers is less than 0.72. also, summary(DebtToIncomeRatio) shows there are almost 7300 NAs.
## 99% quantile is less or equal to 0.72
## 99%
## 0.72
## Summary of DebtToIncomeRatio
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## 0.000 0.150 0.220 0.259 0.320 10.010 7296
Homeowner plot
Inquires last 6 months, most borrowers (99%) have less than 6 inquires. summary(InquiriesLast6Months) shows the whole plot is actually highly right-skewed.
## 99% quantile is less or equal to 6
## 99%
## 6
## Summary of InquiriesLast6Months
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0000 0.0000 0.0000 0.9646 1.0000 27.0000
Trades never delinquent, almost 42% of the borrowers never delinquent. summary(TradesNeverDelinquent) shows ths plot is left-skewed, however, due to the minimum value is 0, the mean is not far away from median.
## 42% borrowers who never delinquent
## [1] 0.4186299
## Summary of TradesNeverDelinquent..percentage.
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0800 0.8500 0.9500 0.9059 1.0000 1.0000
Credit score range lower and upper, for the same borrower, the difference between upper and lower is 19.
##
## 619 639 659 679 699 719 739 759 779 799 819
## 600 1040 0 0 0 0 0 0 0 0 0 0
## 620 0 1653 0 0 0 0 0 0 0 0 0
## 640 0 0 8849 0 0 0 0 0 0 0 0
## 660 0 0 0 14133 0 0 0 0 0 0 0
## 680 0 0 0 0 14019 0 0 0 0 0 0
## 700 0 0 0 0 0 13610 0 0 0 0 0
## 720 0 0 0 0 0 0 11034 0 0 0 0
## 740 0 0 0 0 0 0 0 7871 0 0 0
## 760 0 0 0 0 0 0 0 0 5253 0 0
## 780 0 0 0 0 0 0 0 0 0 3705 0
## 800 0 0 0 0 0 0 0 0 0 0 2107
## 820 0 0 0 0 0 0 0 0 0 0 0
## 840 0 0 0 0 0 0 0 0 0 0 0
## 860 0 0 0 0 0 0 0 0 0 0 0
## 880 0 0 0 0 0 0 0 0 0 0 0
##
## 839 859 879 899
## 600 0 0 0 0
## 620 0 0 0 0
## 640 0 0 0 0
## 660 0 0 0 0
## 680 0 0 0 0
## 700 0 0 0 0
## 720 0 0 0 0
## 740 0 0 0 0
## 760 0 0 0 0
## 780 0 0 0 0
## 800 0 0 0 0
## 820 1042 0 0 0
## 840 0 398 0 0
## 860 0 0 122 0
## 880 0 0 0 17
Bar chart of prosper score, although all the documentation says that the score range from 1 to 10, however, there are almost 1500 score 11 in the dataset.
Monthly payment on revolving accounts at the time the credit profile was pulled. 99% of the borrowers have less than 2022 open revolving monthly payment.
## 99% borrowers have less than 2022 OpenRevolvingMonthlyPayment
## 99%
## 2021.48
The percentage of available revolving credit that is utilized at the time the credit profile was pulled. There are almost 4000 borrowers whose percentage of available revolving credit is 0, and 622 borrowers whose percentage of available revolving credit greater than 1.
## 3945 borrowers whose BankcardUtilization == 0
## pf$BankcardUtilization == 0: FALSE
## [1] 80908
## --------------------------------------------------------
## pf$BankcardUtilization == 0: TRUE
## [1] 3945
## 622 borrowers whose BankcardUtilization > 1
## pf$BankcardUtilization > 1: FALSE
## [1] 84231
## --------------------------------------------------------
## pf$BankcardUtilization > 1: TRUE
## [1] 622
The number of days delinquent. There are 76445 borrowers whose days delinquent is 0. 2nd plot focus on LoanCurrentDaysDelinquent > 1.
## 76445 borrowers whose days delinquent is 0
## pf$LoanCurrentDaysDelinquent == 0: FALSE
## [1] 8408
## --------------------------------------------------------
## pf$LoanCurrentDaysDelinquent == 0: TRUE
## [1] 76445
The distribution of monthly loan payment, there is one thing I feel interesting, there are 91 borrowers, who make less monthly income than monthly loan payment, and their income is verifiable. 68 out of these 91 borrowers are not even a homeowner.
## borrowers who make less income than payment
## [1] 91
## borrowers who make less income than payment and have no house
## [1] 68
I create a new variable called loanToIncomeRatio = LoanOriginalAmount/StatedMonthlyIncome, then I cap it at 19 because 99% of the loanToIncomeRatio are less or equal to 18 and call this new variable loanToIncomeRatioC. Almost 84% of the borrowers borrow loans amount less than 3 months of income.
## 99% quantile is less or equal to 18
## 99%
## 18
## 84% of loanToIncomeRatioC is less or equal to 3
## [1] 0.839275
bar chart of ProsperRating..Alpha.
The original dataset has 113917 listing and 81 features, however, after July 2009, prosper implemented new rating system, therefore, I only use data after July 2009, which is about 75% of the original dataset.
The main feature is LenderYield, I’d like to find out which features are best for predicting lender yield.
Income, property owner, current debt, records of default, purpose of borrowing, credit rating.
I create a new variable called loanToIncomeRatio = LoanOriginalAmount/StatedMonthlyIncome, then I cap it at 19 and call it loanToIncomeRatioC, because 99% of the loanToIncomeRatio are less or equal to 18. And I cap CurrentDelinquencies and InquiriesLast6Months and call them CurrentDelinquenciesF and InquiriesLast6MonthsF for plotting purpose.
BankcardUtilization, this m-shape histogram reminds me disappearing middle class. In the plot, many people either has no money can borrow or already maxed out their credit.
The reason I exclude BorrowerRate and BorrowerAPR these two features is that common sense tells us that if borrowing rate is high, then lender yield is high, therefore, they should have high correlation. And we can also deduce from summary(BorrowerRate - LenderYield) that prosper take 1% of the loans as a fee (40 listing are 2%).
## summary of (BorrowerRate - LenderYield)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.01 0.01 0.01 0.01 0.01 0.02
Higher income, lower lender yield. red x represents mean.
LenderYield against DebtToIncomeRatio, 2nd plot focus on 99% of the DebtToIncomeRatio (0 to 0.72). Blue dotted line are 10% or 90% percentile, blue solid line is median, red line is mean. This is a very intuitive plot, most people wouldn’t consider lending money to a person who have mountain high of debt unless there are enough risk premium. Though this is a very noisy plot, it imples that they might only have very weakly correlation.
##
## Pearson's product-moment correlation
##
## data: pf$LenderYield and pf$DebtToIncomeRatio
## t = 35.495, df = 77555, p-value < 2.2e-16
## alternative hypothesis: true correlation is greater than 0
## 95 percent confidence interval:
## 0.1206188 1.0000000
## sample estimates:
## cor
## 0.1264351
Most home owner can get slightly lower interest rate, another word, for lenders, its risk is lower. 2nd plot shows that they are not normally distributed, so I perform Mann-Whitney-Wilcoxon test to further support the theory.
##
## Wilcoxon rank sum test with continuity correction
##
## data: LenderYield by IsBorrowerHomeowner
## W = 1032300000, p-value < 2.2e-16
## alternative hypothesis: true location shift is not equal to 0
I cap the top 1% of InquiriesLast6Months at 6 and call it InquiriesLast6MonthsF, then draw a box plot, it’s clear that less inquiries, less lender yield (less risk).
This graph shows a trend that borrowers who have utilized more available revolving credit getting higher interest rate (except for 3945 borrowers who have 0 available revolving credit). 664 of 3945 borrowers who have 0 open revolving account, therefore, their average lender yield is higher.
##
## Pearson's product-moment correlation
##
## data: pf$LenderYield and pf$BankcardUtilization
## t = 71.698, df = 84851, p-value < 2.2e-16
## alternative hypothesis: true correlation is greater than 0
## 95 percent confidence interval:
## 0.2336715 1.0000000
## sample estimates:
## cor
## 0.2390029
## 3945 borrowers who have 0 available revolving credit
## pf$BankcardUtilization == 0: FALSE
## [1] 80908
## --------------------------------------------------------
## pf$BankcardUtilization == 0: TRUE
## [1] 3945
## 664 of 3945 borrowers who have 0 open revolving account
## pf$BankcardUtilization == 0 & pf$OpenRevolvingAccounts == 0: FALSE
## [1] 84189
## --------------------------------------------------------
## pf$BankcardUtilization == 0 & pf$OpenRevolvingAccounts == 0: TRUE
## [1] 664
## Summary of LenderYield index by OpenRevolvingAccounts == 0
## pf$OpenRevolvingAccounts == 0: FALSE
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0300 0.1255 0.1774 0.1853 0.2466 0.3400
## --------------------------------------------------------
## pf$OpenRevolvingAccounts == 0: TRUE
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0850 0.2392 0.2999 0.2707 0.3077 0.3400
This plot shows a weak relationship between TradesNeverDelinquent and LenderYield. From this plot, percentage less than 0.6 are very fluctuated.
##
## Pearson's product-moment correlation
##
## data: pf$TradesNeverDelinquent..percentage. and pf$LenderYield
## t = -79.827, df = 84851, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## -0.2705472 -0.2580303
## sample estimates:
## cor
## -0.2642999
A moderate relationship between CreditScoreRangeLower and LenderYield, however, even if you have a very high CreditScore, it seems you still have lower change getting a high borrower rate.
A moderate relationship between ProsperScore and LenderYield, one interesting though, the mean and median of ProsperScore 5 are higher than ProsperScore 4. Further test supports this observation.
##
## Wilcoxon rank sum test with continuity correction
##
## data: pf$LenderYield[pf$ProsperScore == 4] and pf$LenderYield[pf$ProsperScore == 5]
## W = 60491000, p-value = 0.006519
## alternative hypothesis: true location shift is not equal to 0
LenderYield against MonthlyLoanPayment is a very interesting plot, so many straight lines mean some other variables affecting MonthlyLoanPayment.
2nd plot implies that MonthlyLoanPayment are spread loosely, that’s why the mean and other percentile line are so fluctuated because there are maybe only 1 to 10 (x, y) pairs on one x. One variable comes to mind is LoanOriginalAmount because it has high correlation with MonthlyLoanPayment.
So the 3rd plot proves my intuitive, however, there must be another variables affecting both of them, my guess is Term, since there are only 3 terms on prosper loans, and the plot also has 3 ‘lines’.
The 4th plot shows higher the LoanOriginalAmount, lower the LenderYield, it should be because only borrowers have higher rating get to borrow higher amount, and borrowers have higher rating can enjoy lower interest.
One thing strange do occur to me when I look at 3rd plot, there are 484 borrowers whose MonthlyLoanPayment == 0.
## Summary of LoanOriginalAmount indexed by ProsperRating
## pf$ProsperRating..Alpha.: AA
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 6000 10940 11584 16000 35000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: A
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 5850 10000 11460 15000 35000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: B
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 6000 10000 11622 15000 35000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: C
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 5000 10000 10392 15000 25000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: D
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 4000 6100 7083 10000 15000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: E
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 3600 4000 4586 5000 15900
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: HR
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 3000 4000 3463 4000 16800
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: NA
## NULL
## 484 of borrowers whose MonthlyLoanPayment equal 0.
## pf$MonthlyLoanPayment == 0: FALSE
## [1] 84369
## --------------------------------------------------------
## pf$MonthlyLoanPayment == 0: TRUE
## [1] 484
looks like we have found the most important factors deciding lender yield, even with some outliers, the mean still wasn’t affected in all rating categories.
LenderYield strongly correlates with ProsperRating..Alpha., in fact, using linear regression, ProsperRating..Alpha. one variable alone can explain about 91% of the variance of the LenderYield. No other variable can provide such high r-squared.
For many cases, we can all observe a trend, you have a house, less inquiries, high income, lower debtToIncomeRatio, less BankcardUtilization, better credit score can help borrower get a lower interest, especially credit score.
LenderYield against MonthlyLoanPayment is the most interesting plot I have seen so far. This plot looks like somebody randomly draw some line on it.
MonthlyLoanPayment and LoanOriginalAmount, they are highly positively correlated with each other, however, there are 3 lines of them instead of one line in most highly correlated cases.
ProsperRating and LenderYield, their correlation is about -0.95. Basically, if a borrower have higher ProsperRating, he can get a lower borrowing interest.
This density plot reaffirms the boxplot of LenderYield by IncomeRange, borrowers who have higher income can get better interest rate. Although their range still remain the same from 0 to 0.34, the mode shifts.
## Summary of LenderYield indexed by IncomeRange
## pf$IncomeRange: Not employed
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0700 0.2048 0.2759 0.2514 0.3077 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $0
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0820 0.2187 0.2770 0.2555 0.3049 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $1-24,999
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0450 0.1724 0.2398 0.2275 0.3023 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $25,000-49,999
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0398 0.1459 0.2000 0.2037 0.2599 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $50,000-74,999
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0300 0.1255 0.1740 0.1826 0.2399 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $75,000-99,999
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0399 0.1149 0.1634 0.1731 0.2259 0.3400
## --------------------------------------------------------
## pf$IncomeRange: $100,000+
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0350 0.1053 0.1459 0.1602 0.2039 0.3400
## Summary of LenderYield indexed by InquiriesLast6MonthsF
## pf$InquiriesLast6MonthsF: 0
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0323 0.1139 0.1585 0.1695 0.2204 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 1
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0300 0.1340 0.1840 0.1915 0.2499 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 2
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0399 0.1459 0.2024 0.2054 0.2686 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 3
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0325 0.1560 0.2187 0.2152 0.2800 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 4
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0399 0.1700 0.2387 0.2263 0.2958 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 5
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0529 0.1805 0.2473 0.2359 0.3027 0.3400
## --------------------------------------------------------
## pf$InquiriesLast6MonthsF: 6
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0528 0.2099 0.2816 0.2550 0.3077 0.3400
From the density plot of LenderYield by ProsperRating, we can observe that most of them are not overlapping with each other. The worst rating put most weight around 0.3.
## Summary of LenderYield indexed by ProsperRating
## pf$ProsperRating..Alpha.: AA
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.03000 0.05990 0.06790 0.06911 0.07450 0.20000
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: A
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0398 0.0890 0.1019 0.1029 0.1139 0.2050
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: B
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0593 0.1314 0.1409 0.1444 0.1539 0.3400
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: C
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0795 0.1665 0.1814 0.1844 0.1999 0.3400
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: D
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.1057 0.2187 0.2392 0.2364 0.2525 0.3400
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: E
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.1379 0.2612 0.2825 0.2833 0.3049 0.3400
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: HR
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.1679 0.3034 0.3077 0.3073 0.3077 0.3400
## --------------------------------------------------------
## pf$ProsperRating..Alpha.: NA
## NULL
This plot explains why there are so many straight lines in plot of LenderYield against MonthlyLoanPayment. Because under the same loan amount, when LenderYield (borrower interest rate) increases, MonthlyLoanPayment also increase linearly. Loan amount and interest rate decides monthly payment.
ProsperScore has the 2nd highest correlation with LenderYield (-0.65), however, from this plot, we can see that its predicting power is nothing comparing to ProsperRating. Even though with a lower ProsperScore as 3, its average lenderYield still ranges from 0.1 to 0.3.
##
## Call:
## lm(formula = LenderYield ~ ProsperRating..Alpha., data = pf)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.145435 -0.014014 -0.000527 0.011967 0.195551
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.0691108 0.0002989 231.24 <2e-16 ***
## ProsperRating..Alpha.A 0.0338226 0.0003497 96.71 <2e-16 ***
## ProsperRating..Alpha.B 0.0753379 0.0003466 217.37 <2e-16 ***
## ProsperRating..Alpha.C 0.1153158 0.0003398 339.33 <2e-16 ***
## ProsperRating..Alpha.D 0.1673034 0.0003506 477.14 <2e-16 ***
## ProsperRating..Alpha.E 0.2142246 0.0003719 576.01 <2e-16 ***
## ProsperRating..Alpha.HR 0.2382070 0.0003981 598.29 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.02191 on 84846 degrees of freedom
## Multiple R-squared: 0.9139, Adjusted R-squared: 0.9138
## F-statistic: 1.5e+05 on 6 and 84846 DF, p-value: < 2.2e-16
##
## Call:
## lm(formula = ProsperRating..numeric. ~ . - LenderYield, data = pfLM)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.9068 -0.5861 0.0556 0.6466 4.0447
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -5.712e+00 6.396e-02 -89.307 < 2e-16
## IncomeRangeN 2.993e-02 3.611e-03 8.289 < 2e-16
## DebtToIncomeRatio -2.365e-01 1.099e-02 -21.524 < 2e-16
## IsBorrowerHomeownerN -5.136e-02 7.409e-03 -6.932 4.17e-12
## InquiriesLast6Months -8.204e-02 2.651e-03 -30.942 < 2e-16
## TradesNeverDelinquent..percentage. 6.759e-01 3.167e-02 21.343 < 2e-16
## CreditScoreRangeLower 1.016e-02 9.729e-05 104.451 < 2e-16
## ProsperScore 3.457e-01 1.721e-03 200.933 < 2e-16
## OpenRevolvingMonthlyPayment -1.522e-04 9.568e-06 -15.905 < 2e-16
## BankcardUtilization -9.050e-02 1.412e-02 -6.410 1.46e-10
## LoanCurrentDaysDelinquent -1.098e-03 2.337e-05 -46.963 < 2e-16
## MonthlyLoanPayment -3.876e-04 4.507e-05 -8.600 < 2e-16
## LP_InterestandFees -2.722e-04 3.073e-06 -88.569 < 2e-16
## LoanOriginalAmount 7.020e-05 1.384e-06 50.718 < 2e-16
##
## (Intercept) ***
## IncomeRangeN ***
## DebtToIncomeRatio ***
## IsBorrowerHomeownerN ***
## InquiriesLast6Months ***
## TradesNeverDelinquent..percentage. ***
## CreditScoreRangeLower ***
## ProsperScore ***
## OpenRevolvingMonthlyPayment ***
## BankcardUtilization ***
## LoanCurrentDaysDelinquent ***
## MonthlyLoanPayment ***
## LP_InterestandFees ***
## LoanOriginalAmount ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9209 on 77543 degrees of freedom
## (7296 observations deleted due to missingness)
## Multiple R-squared: 0.6918, Adjusted R-squared: 0.6918
## F-statistic: 1.339e+04 on 13 and 77543 DF, p-value: < 2.2e-16
## Start: AIC=-28466.67
## ProsperRating..numeric. ~ (Term + LenderYield + ProsperScore +
## ListingCategory..numeric. + EmploymentStatusDuration + IsBorrowerHomeowner +
## CreditScoreRangeLower + CreditScoreRangeUpper + CurrentCreditLines +
## OpenCreditLines + TotalCreditLinespast7years + OpenRevolvingAccounts +
## OpenRevolvingMonthlyPayment + InquiriesLast6Months + TotalInquiries +
## CurrentDelinquencies + AmountDelinquent + DelinquenciesLast7Years +
## PublicRecordsLast10Years + PublicRecordsLast12Months + RevolvingCreditBalance +
## BankcardUtilization + AvailableBankcardCredit + TotalTrades +
## TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## LoanOriginalAmount + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_InterestandFees + LP_ServiceFees +
## LP_CollectionFees + LP_GrossPrincipalLoss + LP_NetPrincipalLoss +
## LP_NonPrincipalRecoverypayments + PercentFunded + Recommendations +
## InvestmentFromFriendsCount + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC) -
## LenderYield
##
##
## Step: AIC=-28466.67
## ProsperRating..numeric. ~ Term + ProsperScore + ListingCategory..numeric. +
## EmploymentStatusDuration + IsBorrowerHomeowner + CreditScoreRangeLower +
## CreditScoreRangeUpper + CurrentCreditLines + OpenCreditLines +
## TotalCreditLinespast7years + OpenRevolvingAccounts + OpenRevolvingMonthlyPayment +
## InquiriesLast6Months + TotalInquiries + CurrentDelinquencies +
## AmountDelinquent + DelinquenciesLast7Years + PublicRecordsLast10Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## TradesOpenedLast6Months + DebtToIncomeRatio + IncomeRange +
## IncomeVerifiable + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + LoanOriginalAmount + MonthlyLoanPayment +
## LP_CustomerPayments + LP_CustomerPrincipalPayments + LP_ServiceFees +
## LP_CollectionFees + LP_GrossPrincipalLoss + LP_NetPrincipalLoss +
## LP_NonPrincipalRecoverypayments + PercentFunded + Recommendations +
## InvestmentFromFriendsCount + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC
##
##
## Step: AIC=-28466.67
## ProsperRating..numeric. ~ Term + ProsperScore + ListingCategory..numeric. +
## EmploymentStatusDuration + IsBorrowerHomeowner + CreditScoreRangeLower +
## CurrentCreditLines + OpenCreditLines + TotalCreditLinespast7years +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + AmountDelinquent +
## DelinquenciesLast7Years + PublicRecordsLast10Years + PublicRecordsLast12Months +
## RevolvingCreditBalance + BankcardUtilization + AvailableBankcardCredit +
## TotalTrades + TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## LoanOriginalAmount + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + InvestmentFromFriendsCount +
## InvestmentFromFriendsAmount + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
##
## Step: AIC=-28466.67
## ProsperRating..numeric. ~ Term + ProsperScore + EmploymentStatusDuration +
## IsBorrowerHomeowner + CreditScoreRangeLower + CurrentCreditLines +
## OpenCreditLines + TotalCreditLinespast7years + OpenRevolvingAccounts +
## OpenRevolvingMonthlyPayment + InquiriesLast6Months + TotalInquiries +
## CurrentDelinquencies + AmountDelinquent + DelinquenciesLast7Years +
## PublicRecordsLast10Years + PublicRecordsLast12Months + RevolvingCreditBalance +
## BankcardUtilization + AvailableBankcardCredit + TotalTrades +
## TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## LoanOriginalAmount + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + InvestmentFromFriendsCount +
## InvestmentFromFriendsAmount + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - LoanOriginalAmount 1 0.0 53621 -28468.7
## - TotalCreditLinespast7years 1 0.0 53621 -28468.6
## - PublicRecordsLast10Years 1 0.2 53621 -28468.4
## - EmploymentStatusDuration 1 0.2 53621 -28468.3
## - AmountDelinquent 1 0.4 53622 -28468.1
## - InvestmentFromFriendsCount 1 0.6 53622 -28467.8
## - InvestmentFromFriendsAmount 1 1.0 53622 -28467.2
## - TradesOpenedLast6Months 1 1.1 53622 -28467.1
## - IncomeVerifiable 1 1.1 53622 -28467.0
## <none> 53621 -28466.7
## - TotalInquiries 1 2.2 53623 -28465.5
## - PublicRecordsLast12Months 1 2.5 53624 -28465.1
## - CurrentDelinquencies 1 2.7 53624 -28464.8
## - CurrentCreditLines 1 2.8 53624 -28464.7
## - DelinquenciesLast7Years 1 3.5 53625 -28463.6
## - StatedMonthlyIncome 1 4.4 53626 -28462.4
## - TotalTrades 1 4.6 53626 -28462.1
## - OpenCreditLines 1 8.3 53629 -28456.7
## - IsBorrowerHomeowner 1 9.2 53630 -28455.4
## - PercentFunded 1 10.9 53632 -28452.9
## - RevolvingCreditBalance 1 13.5 53635 -28449.1
## - AvailableBankcardCredit 1 14.6 53636 -28447.6
## - OpenRevolvingAccounts 1 15.7 53637 -28446.0
## - Recommendations 1 18.9 53640 -28441.3
## - MonthlyLoanPayment 1 24.3 53645 -28433.6
## - LP_NonPrincipalRecoverypayments 1 32.9 53654 -28421.2
## - Investors 1 40.2 53661 -28410.5
## - LP_GrossPrincipalLoss 1 43.4 53665 -28406.0
## - BankcardUtilization 1 44.6 53666 -28404.2
## - LP_CollectionFees 1 51.9 53673 -28393.7
## - LP_NetPrincipalLoss 1 52.2 53673 -28393.3
## - loanToIncomeRatio 1 89.6 53711 -28339.3
## - OpenRevolvingMonthlyPayment 1 90.0 53711 -28338.6
## - TradesNeverDelinquent..percentage. 1 161.5 53783 -28235.4
## - InquiriesLast6Months 1 168.8 53790 -28224.9
## - Term 1 203.8 53825 -28174.5
## - LoanCurrentDaysDelinquent 1 301.1 53922 -28034.4
## - ListingCategoryF 19 480.4 54102 -27813.1
## - IncomeRange 5 679.4 54301 -27500.3
## - loanToIncomeRatioC 1 1010.0 54631 -27021.7
## - DebtToIncomeRatio 1 1129.2 54750 -26852.6
## - LoanMonthsSinceOrigination 1 2603.8 56225 -24791.8
## - CreditScoreRangeLower 1 5017.2 58638 -21532.9
## - LP_ServiceFees 1 5789.4 59411 -20518.3
## - LP_CustomerPrincipalPayments 1 6735.3 60357 -19293.4
## - LP_CustomerPayments 1 6828.8 60450 -19173.5
## - ProsperScore 1 26817.1 80438 2978.5
##
## Step: AIC=-28468.66
## ProsperRating..numeric. ~ Term + ProsperScore + EmploymentStatusDuration +
## IsBorrowerHomeowner + CreditScoreRangeLower + CurrentCreditLines +
## OpenCreditLines + TotalCreditLinespast7years + OpenRevolvingAccounts +
## OpenRevolvingMonthlyPayment + InquiriesLast6Months + TotalInquiries +
## CurrentDelinquencies + AmountDelinquent + DelinquenciesLast7Years +
## PublicRecordsLast10Years + PublicRecordsLast12Months + RevolvingCreditBalance +
## BankcardUtilization + AvailableBankcardCredit + TotalTrades +
## TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## MonthlyLoanPayment + LP_CustomerPayments + LP_CustomerPrincipalPayments +
## LP_ServiceFees + LP_CollectionFees + LP_GrossPrincipalLoss +
## LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments + PercentFunded +
## Recommendations + InvestmentFromFriendsCount + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - TotalCreditLinespast7years 1 0.0 53621 -28470.6
## - PublicRecordsLast10Years 1 0.2 53621 -28470.4
## - EmploymentStatusDuration 1 0.2 53621 -28470.3
## - AmountDelinquent 1 0.4 53622 -28470.1
## - InvestmentFromFriendsCount 1 0.6 53622 -28469.7
## - InvestmentFromFriendsAmount 1 1.0 53622 -28469.2
## - TradesOpenedLast6Months 1 1.1 53622 -28469.1
## - IncomeVerifiable 1 1.1 53622 -28469.0
## <none> 53621 -28468.7
## - TotalInquiries 1 2.2 53623 -28467.5
## - PublicRecordsLast12Months 1 2.5 53624 -28467.1
## - CurrentDelinquencies 1 2.7 53624 -28466.8
## - CurrentCreditLines 1 2.8 53624 -28466.7
## - DelinquenciesLast7Years 1 3.5 53625 -28465.6
## - StatedMonthlyIncome 1 4.4 53626 -28464.3
## - TotalTrades 1 4.6 53626 -28464.1
## - OpenCreditLines 1 8.3 53630 -28458.7
## - IsBorrowerHomeowner 1 9.2 53630 -28457.4
## - PercentFunded 1 10.9 53632 -28454.9
## - RevolvingCreditBalance 1 13.5 53635 -28451.1
## - AvailableBankcardCredit 1 14.6 53636 -28449.6
## - OpenRevolvingAccounts 1 15.7 53637 -28447.9
## - Recommendations 1 18.9 53640 -28443.3
## - LP_NonPrincipalRecoverypayments 1 32.9 53654 -28423.1
## - Investors 1 40.3 53662 -28412.4
## - LP_GrossPrincipalLoss 1 43.4 53665 -28407.9
## - BankcardUtilization 1 44.6 53666 -28406.2
## - LP_CollectionFees 1 52.0 53673 -28395.5
## - LP_NetPrincipalLoss 1 52.2 53673 -28395.2
## - loanToIncomeRatio 1 89.6 53711 -28341.2
## - OpenRevolvingMonthlyPayment 1 90.1 53711 -28340.5
## - MonthlyLoanPayment 1 105.5 53727 -28318.3
## - TradesNeverDelinquent..percentage. 1 161.6 53783 -28237.4
## - InquiriesLast6Months 1 168.8 53790 -28226.9
## - LoanCurrentDaysDelinquent 1 301.1 53922 -28036.4
## - Term 1 355.5 53977 -27958.3
## - ListingCategoryF 19 482.4 54104 -27812.1
## - IncomeRange 5 693.4 54315 -27482.3
## - loanToIncomeRatioC 1 1077.3 54698 -26928.3
## - DebtToIncomeRatio 1 1158.7 54780 -26812.9
## - LoanMonthsSinceOrigination 1 2609.0 56230 -24786.6
## - CreditScoreRangeLower 1 5026.2 58647 -21523.0
## - LP_ServiceFees 1 6210.8 59832 -19972.3
## - LP_CustomerPrincipalPayments 1 7103.4 60725 -18824.0
## - LP_CustomerPayments 1 7288.8 60910 -18587.6
## - ProsperScore 1 27772.1 81393 3891.8
##
## Step: AIC=-28470.63
## ProsperRating..numeric. ~ Term + ProsperScore + EmploymentStatusDuration +
## IsBorrowerHomeowner + CreditScoreRangeLower + CurrentCreditLines +
## OpenCreditLines + OpenRevolvingAccounts + OpenRevolvingMonthlyPayment +
## InquiriesLast6Months + TotalInquiries + CurrentDelinquencies +
## AmountDelinquent + DelinquenciesLast7Years + PublicRecordsLast10Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## TradesOpenedLast6Months + DebtToIncomeRatio + IncomeRange +
## IncomeVerifiable + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + InvestmentFromFriendsCount +
## InvestmentFromFriendsAmount + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - PublicRecordsLast10Years 1 0.2 53621 -28472
## - EmploymentStatusDuration 1 0.2 53621 -28472
## - AmountDelinquent 1 0.4 53622 -28472
## - InvestmentFromFriendsCount 1 0.6 53622 -28472
## - InvestmentFromFriendsAmount 1 1.0 53622 -28471
## - TradesOpenedLast6Months 1 1.1 53622 -28471
## - IncomeVerifiable 1 1.1 53622 -28471
## <none> 53621 -28471
## - TotalInquiries 1 2.2 53623 -28470
## - PublicRecordsLast12Months 1 2.5 53624 -28469
## - CurrentDelinquencies 1 2.6 53624 -28469
## - CurrentCreditLines 1 2.8 53624 -28469
## - DelinquenciesLast7Years 1 3.5 53625 -28468
## - StatedMonthlyIncome 1 4.4 53626 -28466
## - OpenCreditLines 1 8.4 53630 -28461
## - IsBorrowerHomeowner 1 9.2 53630 -28459
## - PercentFunded 1 10.9 53632 -28457
## - RevolvingCreditBalance 1 13.5 53635 -28453
## - AvailableBankcardCredit 1 14.6 53636 -28452
## - OpenRevolvingAccounts 1 16.2 53637 -28449
## - Recommendations 1 18.9 53640 -28445
## - TotalTrades 1 22.7 53644 -28440
## - LP_NonPrincipalRecoverypayments 1 32.9 53654 -28425
## - Investors 1 40.3 53662 -28414
## - LP_GrossPrincipalLoss 1 43.5 53665 -28410
## - BankcardUtilization 1 44.6 53666 -28408
## - LP_CollectionFees 1 52.0 53673 -28398
## - LP_NetPrincipalLoss 1 52.2 53673 -28397
## - loanToIncomeRatio 1 89.6 53711 -28343
## - OpenRevolvingMonthlyPayment 1 90.2 53711 -28342
## - MonthlyLoanPayment 1 105.5 53727 -28320
## - TradesNeverDelinquent..percentage. 1 161.7 53783 -28239
## - InquiriesLast6Months 1 168.8 53790 -28229
## - LoanCurrentDaysDelinquent 1 301.1 53922 -28038
## - Term 1 355.5 53977 -27960
## - ListingCategoryF 19 482.4 54104 -27814
## - IncomeRange 5 693.4 54315 -27484
## - loanToIncomeRatioC 1 1077.5 54699 -26930
## - DebtToIncomeRatio 1 1158.9 54780 -26815
## - LoanMonthsSinceOrigination 1 2609.0 56230 -24789
## - CreditScoreRangeLower 1 5031.2 58652 -21518
## - LP_ServiceFees 1 6211.3 59833 -19974
## - LP_CustomerPrincipalPayments 1 7104.1 60725 -18825
## - LP_CustomerPayments 1 7289.7 60911 -18588
## - ProsperScore 1 27774.4 81396 3892
##
## Step: AIC=-28472.39
## ProsperRating..numeric. ~ Term + ProsperScore + EmploymentStatusDuration +
## IsBorrowerHomeowner + CreditScoreRangeLower + CurrentCreditLines +
## OpenCreditLines + OpenRevolvingAccounts + OpenRevolvingMonthlyPayment +
## InquiriesLast6Months + TotalInquiries + CurrentDelinquencies +
## AmountDelinquent + DelinquenciesLast7Years + PublicRecordsLast12Months +
## RevolvingCreditBalance + BankcardUtilization + AvailableBankcardCredit +
## TotalTrades + TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## MonthlyLoanPayment + LP_CustomerPayments + LP_CustomerPrincipalPayments +
## LP_ServiceFees + LP_CollectionFees + LP_GrossPrincipalLoss +
## LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments + PercentFunded +
## Recommendations + InvestmentFromFriendsCount + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - EmploymentStatusDuration 1 0.2 53622 -28474.1
## - AmountDelinquent 1 0.4 53622 -28473.8
## - InvestmentFromFriendsCount 1 0.6 53622 -28473.5
## - InvestmentFromFriendsAmount 1 1.0 53622 -28473.0
## - TradesOpenedLast6Months 1 1.1 53623 -28472.8
## - IncomeVerifiable 1 1.1 53623 -28472.8
## <none> 53621 -28472.4
## - TotalInquiries 1 2.1 53624 -28471.3
## - CurrentDelinquencies 1 2.7 53624 -28470.5
## - PublicRecordsLast12Months 1 2.9 53624 -28470.1
## - CurrentCreditLines 1 3.0 53624 -28470.1
## - DelinquenciesLast7Years 1 3.6 53625 -28469.2
## - StatedMonthlyIncome 1 4.4 53626 -28468.1
## - OpenCreditLines 1 8.3 53630 -28462.4
## - IsBorrowerHomeowner 1 9.1 53631 -28461.2
## - PercentFunded 1 10.9 53632 -28458.6
## - RevolvingCreditBalance 1 13.5 53635 -28454.8
## - AvailableBankcardCredit 1 14.5 53636 -28453.4
## - OpenRevolvingAccounts 1 16.3 53638 -28450.8
## - Recommendations 1 18.9 53640 -28447.0
## - TotalTrades 1 22.6 53644 -28441.8
## - LP_NonPrincipalRecoverypayments 1 32.9 53654 -28426.8
## - Investors 1 40.4 53662 -28416.1
## - LP_GrossPrincipalLoss 1 43.5 53665 -28411.6
## - BankcardUtilization 1 44.5 53666 -28410.0
## - LP_CollectionFees 1 52.1 53673 -28399.1
## - LP_NetPrincipalLoss 1 52.2 53674 -28398.9
## - loanToIncomeRatio 1 89.6 53711 -28345.0
## - OpenRevolvingMonthlyPayment 1 90.0 53711 -28344.4
## - MonthlyLoanPayment 1 105.3 53727 -28322.2
## - TradesNeverDelinquent..percentage. 1 164.3 53786 -28237.2
## - InquiriesLast6Months 1 168.7 53790 -28230.8
## - LoanCurrentDaysDelinquent 1 301.1 53923 -28040.2
## - Term 1 355.4 53977 -27962.2
## - ListingCategoryF 19 482.3 54104 -27816.1
## - IncomeRange 5 693.3 54315 -27486.2
## - loanToIncomeRatioC 1 1077.4 54699 -26931.7
## - DebtToIncomeRatio 1 1158.8 54780 -26816.5
## - LoanMonthsSinceOrigination 1 2612.5 56234 -24785.5
## - CreditScoreRangeLower 1 5075.2 58697 -21461.9
## - LP_ServiceFees 1 6212.7 59834 -19973.6
## - LP_CustomerPrincipalPayments 1 7104.2 60726 -18826.7
## - LP_CustomerPayments 1 7289.8 60911 -18590.0
## - ProsperScore 1 27783.9 81405 3899.2
##
## Step: AIC=-28474.09
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + AmountDelinquent +
## DelinquenciesLast7Years + PublicRecordsLast12Months + RevolvingCreditBalance +
## BankcardUtilization + AvailableBankcardCredit + TotalTrades +
## TradesNeverDelinquent..percentage. + TradesOpenedLast6Months +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## MonthlyLoanPayment + LP_CustomerPayments + LP_CustomerPrincipalPayments +
## LP_ServiceFees + LP_CollectionFees + LP_GrossPrincipalLoss +
## LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments + PercentFunded +
## Recommendations + InvestmentFromFriendsCount + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - AmountDelinquent 1 0.4 53622 -28475.5
## - InvestmentFromFriendsCount 1 0.6 53622 -28475.2
## - InvestmentFromFriendsAmount 1 1.0 53623 -28474.7
## - TradesOpenedLast6Months 1 1.1 53623 -28474.5
## - IncomeVerifiable 1 1.1 53623 -28474.4
## <none> 53622 -28474.1
## - TotalInquiries 1 2.1 53624 -28473.1
## - CurrentDelinquencies 1 2.6 53624 -28472.3
## - PublicRecordsLast12Months 1 2.9 53625 -28471.8
## - CurrentCreditLines 1 3.0 53625 -28471.8
## - DelinquenciesLast7Years 1 3.6 53625 -28470.9
## - StatedMonthlyIncome 1 4.4 53626 -28469.8
## - OpenCreditLines 1 8.2 53630 -28464.3
## - IsBorrowerHomeowner 1 8.9 53631 -28463.2
## - PercentFunded 1 10.9 53633 -28460.3
## - RevolvingCreditBalance 1 13.5 53635 -28456.5
## - AvailableBankcardCredit 1 14.6 53636 -28455.0
## - OpenRevolvingAccounts 1 16.1 53638 -28452.8
## - Recommendations 1 18.9 53641 -28448.7
## - TotalTrades 1 23.2 53645 -28442.6
## - LP_NonPrincipalRecoverypayments 1 32.9 53655 -28428.5
## - Investors 1 40.3 53662 -28417.8
## - LP_GrossPrincipalLoss 1 43.4 53665 -28413.3
## - BankcardUtilization 1 44.4 53666 -28411.9
## - LP_CollectionFees 1 52.1 53674 -28400.8
## - LP_NetPrincipalLoss 1 52.2 53674 -28400.6
## - loanToIncomeRatio 1 89.6 53711 -28346.7
## - OpenRevolvingMonthlyPayment 1 89.8 53711 -28346.4
## - MonthlyLoanPayment 1 105.4 53727 -28323.8
## - TradesNeverDelinquent..percentage. 1 164.1 53786 -28239.1
## - InquiriesLast6Months 1 168.8 53790 -28232.4
## - LoanCurrentDaysDelinquent 1 301.1 53923 -28041.9
## - Term 1 355.2 53977 -27964.2
## - ListingCategoryF 19 482.3 54104 -27817.8
## - IncomeRange 5 698.0 54320 -27481.2
## - loanToIncomeRatioC 1 1077.3 54699 -26933.6
## - DebtToIncomeRatio 1 1158.6 54780 -26818.5
## - LoanMonthsSinceOrigination 1 2616.4 56238 -24781.9
## - CreditScoreRangeLower 1 5077.4 58699 -21460.8
## - LP_ServiceFees 1 6212.5 59834 -19975.5
## - LP_CustomerPrincipalPayments 1 7104.4 60726 -18828.2
## - LP_CustomerPayments 1 7289.8 60911 -18591.8
## - ProsperScore 1 27784.3 81406 3897.8
##
## Step: AIC=-28475.49
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + DelinquenciesLast7Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## TradesOpenedLast6Months + DebtToIncomeRatio + IncomeRange +
## IncomeVerifiable + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + InvestmentFromFriendsCount +
## InvestmentFromFriendsAmount + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - InvestmentFromFriendsCount 1 0.6 53623 -28476.6
## - InvestmentFromFriendsAmount 1 1.0 53623 -28476.1
## - TradesOpenedLast6Months 1 1.1 53623 -28475.9
## - IncomeVerifiable 1 1.1 53623 -28475.8
## <none> 53622 -28475.5
## - TotalInquiries 1 2.1 53624 -28474.5
## - CurrentCreditLines 1 3.0 53625 -28473.2
## - PublicRecordsLast12Months 1 3.0 53625 -28473.2
## - CurrentDelinquencies 1 3.6 53626 -28472.2
## - DelinquenciesLast7Years 1 4.0 53626 -28471.7
## - StatedMonthlyIncome 1 4.3 53626 -28471.2
## - OpenCreditLines 1 8.2 53630 -28465.6
## - IsBorrowerHomeowner 1 9.2 53631 -28464.2
## - PercentFunded 1 10.9 53633 -28461.7
## - RevolvingCreditBalance 1 13.5 53636 -28458.0
## - AvailableBankcardCredit 1 14.7 53637 -28456.2
## - OpenRevolvingAccounts 1 16.2 53638 -28454.0
## - Recommendations 1 18.9 53641 -28450.1
## - TotalTrades 1 23.1 53645 -28444.0
## - LP_NonPrincipalRecoverypayments 1 32.9 53655 -28429.9
## - Investors 1 40.3 53662 -28419.2
## - LP_GrossPrincipalLoss 1 43.4 53665 -28414.7
## - BankcardUtilization 1 44.4 53666 -28413.3
## - LP_CollectionFees 1 52.0 53674 -28402.3
## - LP_NetPrincipalLoss 1 52.2 53674 -28402.1
## - loanToIncomeRatio 1 89.6 53712 -28348.1
## - OpenRevolvingMonthlyPayment 1 89.6 53712 -28348.0
## - MonthlyLoanPayment 1 105.3 53727 -28325.3
## - TradesNeverDelinquent..percentage. 1 163.8 53786 -28241.0
## - InquiriesLast6Months 1 168.7 53791 -28234.0
## - LoanCurrentDaysDelinquent 1 301.0 53923 -28043.4
## - Term 1 355.0 53977 -27965.8
## - ListingCategoryF 19 482.5 54104 -27818.9
## - IncomeRange 5 697.7 54320 -27483.1
## - loanToIncomeRatioC 1 1077.3 54699 -26935.1
## - DebtToIncomeRatio 1 1158.5 54781 -26820.0
## - LoanMonthsSinceOrigination 1 2616.1 56238 -24783.8
## - CreditScoreRangeLower 1 5079.4 58701 -21459.5
## - LP_ServiceFees 1 6214.0 59836 -19975.1
## - LP_CustomerPrincipalPayments 1 7106.5 60729 -18826.9
## - LP_CustomerPayments 1 7292.2 60914 -18590.3
## - ProsperScore 1 27805.9 81428 3916.7
##
## Step: AIC=-28476.59
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + DelinquenciesLast7Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## TradesOpenedLast6Months + DebtToIncomeRatio + IncomeRange +
## IncomeVerifiable + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + InvestmentFromFriendsAmount +
## Investors + ListingCategoryF + loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - InvestmentFromFriendsAmount 1 0.5 53623 -28477.8
## - TradesOpenedLast6Months 1 1.1 53624 -28477.0
## - IncomeVerifiable 1 1.1 53624 -28476.9
## <none> 53623 -28476.6
## - TotalInquiries 1 2.1 53625 -28475.6
## - CurrentCreditLines 1 3.0 53626 -28474.3
## - PublicRecordsLast12Months 1 3.0 53626 -28474.3
## - CurrentDelinquencies 1 3.6 53626 -28473.3
## - DelinquenciesLast7Years 1 4.0 53627 -28472.8
## - StatedMonthlyIncome 1 4.3 53627 -28472.3
## - OpenCreditLines 1 8.3 53631 -28466.7
## - IsBorrowerHomeowner 1 9.2 53632 -28465.3
## - PercentFunded 1 10.9 53634 -28462.9
## - RevolvingCreditBalance 1 13.5 53636 -28459.1
## - AvailableBankcardCredit 1 14.6 53637 -28457.4
## - OpenRevolvingAccounts 1 16.2 53639 -28455.1
## - TotalTrades 1 23.1 53646 -28445.2
## - Recommendations 1 28.2 53651 -28437.8
## - LP_NonPrincipalRecoverypayments 1 33.0 53656 -28430.9
## - Investors 1 40.5 53663 -28420.1
## - LP_GrossPrincipalLoss 1 43.4 53666 -28415.8
## - BankcardUtilization 1 44.4 53667 -28414.5
## - LP_CollectionFees 1 52.1 53675 -28403.4
## - LP_NetPrincipalLoss 1 52.2 53675 -28403.1
## - OpenRevolvingMonthlyPayment 1 89.6 53712 -28349.2
## - loanToIncomeRatio 1 89.6 53712 -28349.2
## - MonthlyLoanPayment 1 105.3 53728 -28326.5
## - TradesNeverDelinquent..percentage. 1 163.8 53786 -28242.0
## - InquiriesLast6Months 1 168.8 53791 -28234.9
## - LoanCurrentDaysDelinquent 1 301.4 53924 -28044.0
## - Term 1 354.9 53978 -27967.1
## - ListingCategoryF 19 482.7 54105 -27819.7
## - IncomeRange 5 697.9 54321 -27483.9
## - loanToIncomeRatioC 1 1077.7 54700 -26935.7
## - DebtToIncomeRatio 1 1158.9 54782 -26820.6
## - LoanMonthsSinceOrigination 1 2618.0 56241 -24782.3
## - CreditScoreRangeLower 1 5082.2 58705 -21457.0
## - LP_ServiceFees 1 6213.8 59836 -19976.5
## - LP_CustomerPrincipalPayments 1 7108.9 60732 -18825.2
## - LP_CustomerPayments 1 7294.7 60917 -18588.3
## - ProsperScore 1 27807.8 81430 3917.1
##
## Step: AIC=-28477.81
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + DelinquenciesLast7Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## TradesOpenedLast6Months + DebtToIncomeRatio + IncomeRange +
## IncomeVerifiable + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - TradesOpenedLast6Months 1 1.1 53624 -28478.3
## - IncomeVerifiable 1 1.1 53624 -28478.2
## <none> 53623 -28477.8
## - TotalInquiries 1 2.1 53625 -28476.8
## - CurrentCreditLines 1 3.0 53626 -28475.5
## - PublicRecordsLast12Months 1 3.0 53626 -28475.5
## - CurrentDelinquencies 1 3.6 53627 -28474.6
## - DelinquenciesLast7Years 1 4.0 53627 -28474.0
## - StatedMonthlyIncome 1 4.3 53628 -28473.6
## - OpenCreditLines 1 8.2 53631 -28467.9
## - IsBorrowerHomeowner 1 9.2 53632 -28466.5
## - PercentFunded 1 10.9 53634 -28464.1
## - RevolvingCreditBalance 1 13.5 53637 -28460.3
## - AvailableBankcardCredit 1 14.6 53638 -28458.6
## - OpenRevolvingAccounts 1 16.2 53639 -28456.3
## - TotalTrades 1 23.1 53646 -28446.5
## - Recommendations 1 27.6 53651 -28439.8
## - LP_NonPrincipalRecoverypayments 1 33.1 53656 -28432.0
## - Investors 1 40.7 53664 -28421.0
## - LP_GrossPrincipalLoss 1 43.4 53667 -28417.1
## - BankcardUtilization 1 44.3 53667 -28415.8
## - LP_CollectionFees 1 52.1 53675 -28404.5
## - LP_NetPrincipalLoss 1 52.1 53675 -28404.5
## - OpenRevolvingMonthlyPayment 1 89.6 53713 -28350.4
## - loanToIncomeRatio 1 89.6 53713 -28350.4
## - MonthlyLoanPayment 1 105.5 53729 -28327.4
## - TradesNeverDelinquent..percentage. 1 163.8 53787 -28243.3
## - InquiriesLast6Months 1 168.7 53792 -28236.2
## - LoanCurrentDaysDelinquent 1 301.5 53925 -28045.0
## - Term 1 355.1 53978 -27968.0
## - ListingCategoryF 19 482.6 54106 -27821.0
## - IncomeRange 5 697.8 54321 -27485.3
## - loanToIncomeRatioC 1 1077.5 54701 -26937.1
## - DebtToIncomeRatio 1 1158.7 54782 -26822.1
## - LoanMonthsSinceOrigination 1 2624.9 56248 -24774.0
## - CreditScoreRangeLower 1 5082.1 58705 -21458.4
## - LP_ServiceFees 1 6214.2 59837 -19977.3
## - LP_CustomerPrincipalPayments 1 7108.3 60732 -18827.2
## - LP_CustomerPayments 1 7294.2 60917 -18590.2
## - ProsperScore 1 27818.1 81441 3925.5
##
## Step: AIC=-28478.27
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + DelinquenciesLast7Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## DebtToIncomeRatio + IncomeRange + IncomeVerifiable + StatedMonthlyIncome +
## LoanCurrentDaysDelinquent + LoanMonthsSinceOrigination +
## MonthlyLoanPayment + LP_CustomerPayments + LP_CustomerPrincipalPayments +
## LP_ServiceFees + LP_CollectionFees + LP_GrossPrincipalLoss +
## LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments + PercentFunded +
## Recommendations + Investors + ListingCategoryF + loanToIncomeRatio +
## loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## - IncomeVerifiable 1 1.2 53625 -28478.6
## <none> 53624 -28478.3
## - TotalInquiries 1 2.3 53627 -28477.0
## - CurrentCreditLines 1 2.7 53627 -28476.4
## - PublicRecordsLast12Months 1 3.0 53627 -28476.0
## - CurrentDelinquencies 1 3.4 53628 -28475.3
## - DelinquenciesLast7Years 1 4.1 53628 -28474.3
## - StatedMonthlyIncome 1 4.3 53629 -28474.1
## - OpenCreditLines 1 8.5 53633 -28468.0
## - IsBorrowerHomeowner 1 9.3 53634 -28466.8
## - PercentFunded 1 10.8 53635 -28464.6
## - RevolvingCreditBalance 1 13.5 53638 -28460.8
## - AvailableBankcardCredit 1 14.9 53639 -28458.7
## - OpenRevolvingAccounts 1 16.6 53641 -28456.3
## - TotalTrades 1 22.7 53647 -28447.5
## - Recommendations 1 27.8 53652 -28440.1
## - LP_NonPrincipalRecoverypayments 1 33.1 53657 -28432.4
## - Investors 1 40.7 53665 -28421.5
## - BankcardUtilization 1 43.3 53668 -28417.7
## - LP_GrossPrincipalLoss 1 43.4 53668 -28417.5
## - LP_CollectionFees 1 52.1 53676 -28404.9
## - LP_NetPrincipalLoss 1 52.2 53676 -28404.9
## - OpenRevolvingMonthlyPayment 1 88.5 53713 -28352.4
## - loanToIncomeRatio 1 89.3 53714 -28351.3
## - MonthlyLoanPayment 1 105.5 53730 -28327.9
## - TradesNeverDelinquent..percentage. 1 162.8 53787 -28245.2
## - InquiriesLast6Months 1 181.5 53806 -28218.3
## - LoanCurrentDaysDelinquent 1 302.0 53926 -28044.8
## - Term 1 355.6 53980 -27967.7
## - ListingCategoryF 19 482.7 54107 -27821.4
## - IncomeRange 5 696.7 54321 -27487.3
## - loanToIncomeRatioC 1 1078.2 54702 -26936.7
## - DebtToIncomeRatio 1 1158.1 54782 -26823.5
## - LoanMonthsSinceOrigination 1 2626.1 56250 -24772.8
## - CreditScoreRangeLower 1 5122.2 58746 -21406.1
## - LP_ServiceFees 1 6220.3 59845 -19970.0
## - LP_CustomerPrincipalPayments 1 7111.9 60736 -18823.3
## - LP_CustomerPayments 1 7297.8 60922 -18586.3
## - ProsperScore 1 28655.3 82280 4717.5
##
## Step: AIC=-28478.58
## ProsperRating..numeric. ~ Term + ProsperScore + IsBorrowerHomeowner +
## CreditScoreRangeLower + CurrentCreditLines + OpenCreditLines +
## OpenRevolvingAccounts + OpenRevolvingMonthlyPayment + InquiriesLast6Months +
## TotalInquiries + CurrentDelinquencies + DelinquenciesLast7Years +
## PublicRecordsLast12Months + RevolvingCreditBalance + BankcardUtilization +
## AvailableBankcardCredit + TotalTrades + TradesNeverDelinquent..percentage. +
## DebtToIncomeRatio + IncomeRange + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC
##
## Df Sum of Sq RSS AIC
## <none> 53625 -28478.6
## - TotalInquiries 1 2.3 53628 -28477.3
## - CurrentCreditLines 1 2.7 53628 -28476.7
## - PublicRecordsLast12Months 1 3.0 53628 -28476.3
## - CurrentDelinquencies 1 3.5 53629 -28475.6
## - DelinquenciesLast7Years 1 4.1 53630 -28474.6
## - StatedMonthlyIncome 1 4.3 53630 -28474.3
## - OpenCreditLines 1 8.5 53634 -28468.3
## - IsBorrowerHomeowner 1 9.3 53635 -28467.1
## - PercentFunded 1 10.8 53636 -28464.9
## - RevolvingCreditBalance 1 13.4 53639 -28461.1
## - AvailableBankcardCredit 1 14.9 53640 -28459.1
## - OpenRevolvingAccounts 1 16.5 53642 -28456.7
## - TotalTrades 1 22.6 53648 -28447.8
## - Recommendations 1 27.8 53653 -28440.5
## - LP_NonPrincipalRecoverypayments 1 33.1 53659 -28432.7
## - Investors 1 40.7 53666 -28421.8
## - BankcardUtilization 1 43.3 53669 -28418.0
## - LP_GrossPrincipalLoss 1 43.5 53669 -28417.8
## - LP_CollectionFees 1 52.1 53678 -28405.2
## - LP_NetPrincipalLoss 1 52.2 53678 -28405.2
## - OpenRevolvingMonthlyPayment 1 88.6 53714 -28352.6
## - loanToIncomeRatio 1 89.1 53715 -28351.8
## - MonthlyLoanPayment 1 105.5 53731 -28328.2
## - TradesNeverDelinquent..percentage. 1 162.9 53788 -28245.5
## - InquiriesLast6Months 1 181.7 53807 -28218.3
## - LoanCurrentDaysDelinquent 1 302.1 53927 -28045.0
## - Term 1 355.6 53981 -27968.0
## - ListingCategoryF 19 482.7 54108 -27821.7
## - IncomeRange 5 696.8 54322 -27487.5
## - loanToIncomeRatioC 1 1077.7 54703 -26937.7
## - DebtToIncomeRatio 1 1157.0 54782 -26825.3
## - LoanMonthsSinceOrigination 1 2627.2 56253 -24771.8
## - CreditScoreRangeLower 1 5122.1 58748 -21406.7
## - LP_ServiceFees 1 6222.0 59847 -19968.2
## - LP_CustomerPrincipalPayments 1 7113.5 60739 -18821.7
## - LP_CustomerPayments 1 7299.5 60925 -18584.6
## - ProsperScore 1 28658.1 82284 4719.3
##
## Call:
## lm(formula = ProsperRating..numeric. ~ Term + ProsperScore +
## IsBorrowerHomeowner + CreditScoreRangeLower + CurrentCreditLines +
## OpenCreditLines + OpenRevolvingAccounts + OpenRevolvingMonthlyPayment +
## InquiriesLast6Months + TotalInquiries + CurrentDelinquencies +
## DelinquenciesLast7Years + PublicRecordsLast12Months + RevolvingCreditBalance +
## BankcardUtilization + AvailableBankcardCredit + TotalTrades +
## TradesNeverDelinquent..percentage. + DebtToIncomeRatio +
## IncomeRange + StatedMonthlyIncome + LoanCurrentDaysDelinquent +
## LoanMonthsSinceOrigination + MonthlyLoanPayment + LP_CustomerPayments +
## LP_CustomerPrincipalPayments + LP_ServiceFees + LP_CollectionFees +
## LP_GrossPrincipalLoss + LP_NetPrincipalLoss + LP_NonPrincipalRecoverypayments +
## PercentFunded + Recommendations + Investors + ListingCategoryF +
## loanToIncomeRatio + loanToIncomeRatioC, data = pfLM1.na)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5.0888 -0.5231 0.0294 0.5621 6.0011
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -4.609e+00 8.691e-01 -5.303 1.14e-07
## Term -7.185e-03 3.170e-04 -22.669 < 2e-16
## ProsperScore 3.517e-01 1.729e-03 203.489 < 2e-16
## IsBorrowerHomeownerTRUE -2.566e-02 6.997e-03 -3.668 0.000245
## CreditScoreRangeLower 8.309e-03 9.659e-05 86.028 < 2e-16
## CurrentCreditLines 4.412e-03 2.242e-03 1.968 0.049092
## OpenCreditLines 8.976e-03 2.565e-03 3.499 0.000467
## OpenRevolvingAccounts -7.483e-03 1.532e-03 -4.884 1.04e-06
## OpenRevolvingMonthlyPayment -1.534e-04 1.356e-05 -11.313 < 2e-16
## InquiriesLast6Months -4.526e-02 2.793e-03 -16.204 < 2e-16
## TotalInquiries 1.830e-03 1.005e-03 1.821 0.068579
## CurrentDelinquencies -6.449e-03 2.888e-03 -2.233 0.025520
## DelinquenciesLast7Years -9.637e-04 3.943e-04 -2.444 0.014514
## PublicRecordsLast12Months -6.026e-02 2.912e-02 -2.069 0.038510
## RevolvingCreditBalance 6.629e-07 1.504e-07 4.407 1.05e-05
## BankcardUtilization -1.095e-01 1.384e-02 -7.908 2.65e-15
## AvailableBankcardCredit -1.033e-06 2.229e-07 -4.632 3.62e-06
## TotalTrades 2.187e-03 3.824e-04 5.719 1.07e-08
## TradesNeverDelinquent..percentage. 5.144e-01 3.353e-02 15.340 < 2e-16
## DebtToIncomeRatio -5.985e-01 1.464e-02 -40.887 < 2e-16
## IncomeRange$1-24,999 -2.523e-01 8.324e-01 -0.303 0.761847
## IncomeRange$25,000-49,999 1.116e-01 8.323e-01 0.134 0.893376
## IncomeRange$50,000-74,999 2.624e-01 8.323e-01 0.315 0.752554
## IncomeRange$75,000-99,999 3.180e-01 8.323e-01 0.382 0.702409
## IncomeRange$100,000+ 3.606e-01 8.324e-01 0.433 0.664875
## StatedMonthlyIncome 1.952e-06 7.810e-07 2.499 0.012455
## LoanCurrentDaysDelinquent -5.816e-04 2.784e-05 -20.891 < 2e-16
## LoanMonthsSinceOrigination -2.308e-02 3.746e-04 -61.611 < 2e-16
## MonthlyLoanPayment -3.605e-04 2.920e-05 -12.347 < 2e-16
## LP_CustomerPayments -7.047e-04 6.862e-06 -102.699 < 2e-16
## LP_CustomerPrincipalPayments 6.937e-04 6.843e-06 101.382 < 2e-16
## LP_ServiceFees -1.312e-02 1.384e-04 -94.817 < 2e-16
## LP_CollectionFees -5.519e-04 6.359e-05 -8.679 < 2e-16
## LP_GrossPrincipalLoss -1.653e-04 2.087e-05 -7.923 2.34e-15
## LP_NetPrincipalLoss 1.814e-04 2.089e-05 8.683 < 2e-16
## LP_NonPrincipalRecoverypayments -2.444e-04 3.532e-05 -6.919 4.59e-12
## PercentFunded 5.840e-01 1.476e-01 3.956 7.62e-05
## Recommendations 1.024e-01 1.616e-02 6.333 2.42e-10
## Investors 3.096e-04 4.039e-05 7.664 1.82e-14
## ListingCategoryFDebt Consolidation 2.518e-01 1.910e-01 1.319 0.187274
## ListingCategoryFHome Improvement 1.197e-01 1.912e-01 0.626 0.531250
## ListingCategoryFBusiness 6.409e-02 1.914e-01 0.335 0.737761
## ListingCategoryFStudent Use 3.862e-01 1.999e-01 1.932 0.053374
## ListingCategoryFAuto 2.062e-01 1.918e-01 1.075 0.282418
## ListingCategoryFOther 1.336e-01 1.911e-01 0.699 0.484385
## ListingCategoryFBaby&Adoption 1.079e-01 2.002e-01 0.539 0.589799
## ListingCategoryFBoat 1.978e-01 2.117e-01 0.935 0.349982
## ListingCategoryFCosmetic Procedure -1.750e-01 2.119e-01 -0.826 0.408979
## ListingCategoryFEngagement Ring -5.957e-02 1.998e-01 -0.298 0.765571
## ListingCategoryFGreen Loans 4.706e-02 2.270e-01 0.207 0.835755
## ListingCategoryFHousehold Expenses -8.124e-02 1.919e-01 -0.423 0.672087
## ListingCategoryFLarge Purchases 1.380e-01 1.932e-01 0.714 0.475121
## ListingCategoryFMedical/Dental 5.809e-02 1.922e-01 0.302 0.762485
## ListingCategoryFMotorcycle 1.066e-01 1.971e-01 0.541 0.588711
## ListingCategoryFRV 2.213e-01 2.243e-01 0.987 0.323837
## ListingCategoryFTaxes 5.954e-02 1.932e-01 0.308 0.757944
## ListingCategoryFVacation 5.787e-02 1.934e-01 0.299 0.764812
## ListingCategoryFWedding Loans 4.077e-02 1.934e-01 0.211 0.833034
## loanToIncomeRatio 5.359e-04 4.723e-05 11.347 < 2e-16
## loanToIncomeRatioC 1.812e-01 4.591e-03 39.461 < 2e-16
##
## (Intercept) ***
## Term ***
## ProsperScore ***
## IsBorrowerHomeownerTRUE ***
## CreditScoreRangeLower ***
## CurrentCreditLines *
## OpenCreditLines ***
## OpenRevolvingAccounts ***
## OpenRevolvingMonthlyPayment ***
## InquiriesLast6Months ***
## TotalInquiries .
## CurrentDelinquencies *
## DelinquenciesLast7Years *
## PublicRecordsLast12Months *
## RevolvingCreditBalance ***
## BankcardUtilization ***
## AvailableBankcardCredit ***
## TotalTrades ***
## TradesNeverDelinquent..percentage. ***
## DebtToIncomeRatio ***
## IncomeRange$1-24,999
## IncomeRange$25,000-49,999
## IncomeRange$50,000-74,999
## IncomeRange$75,000-99,999
## IncomeRange$100,000+
## StatedMonthlyIncome *
## LoanCurrentDaysDelinquent ***
## LoanMonthsSinceOrigination ***
## MonthlyLoanPayment ***
## LP_CustomerPayments ***
## LP_CustomerPrincipalPayments ***
## LP_ServiceFees ***
## LP_CollectionFees ***
## LP_GrossPrincipalLoss ***
## LP_NetPrincipalLoss ***
## LP_NonPrincipalRecoverypayments ***
## PercentFunded ***
## Recommendations ***
## Investors ***
## ListingCategoryFDebt Consolidation
## ListingCategoryFHome Improvement
## ListingCategoryFBusiness
## ListingCategoryFStudent Use .
## ListingCategoryFAuto
## ListingCategoryFOther
## ListingCategoryFBaby&Adoption
## ListingCategoryFBoat
## ListingCategoryFCosmetic Procedure
## ListingCategoryFEngagement Ring
## ListingCategoryFGreen Loans
## ListingCategoryFHousehold Expenses
## ListingCategoryFLarge Purchases
## ListingCategoryFMedical/Dental
## ListingCategoryFMotorcycle
## ListingCategoryFRV
## ListingCategoryFTaxes
## ListingCategoryFVacation
## ListingCategoryFWedding Loans
## loanToIncomeRatio ***
## loanToIncomeRatioC ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.8319 on 77483 degrees of freedom
## Multiple R-squared: 0.7487, Adjusted R-squared: 0.7485
## F-statistic: 3912 on 59 and 77483 DF, p-value: < 2.2e-16
From the density plot, we can observe different level of many variables will shift the mode a little bit, but overall, their range still overlap except ProsperRating. The lenderYield density plot facet wrap by ProsperRating shows a most non-overlapping plot.
Yes, LenderYield by MonthlyLoanPayment and LoanOriginalAmount, at first glance, I know that it must has third variable interact with them, and this variable should have linear relationship with one of them. However, it did take me a while to find out what the variable is.
Yes, I created 3 models, first one is just using formula = LenderYield ~ ProsperRating..Alpha., and its r-squared is 0.9138. A little digging up told me that the borrower rates are actually decided by the ProsperRating. So my interest turns to predict ProsperRating using other variables (of course not including LenderYield and other highly correlated variables), I hand-picked 13 variables to fit a linear model against ProsperRating, and I got 0.6918 r-squared. Lastly, I removed some variable then use step() function to find out a linear model using 34 independent variables to predict ProsperRating, which has 0.74 r-squared.
This m-shaped distribution represents many infamous phenomenon during these era, such as disappearing middle class, poor become poorer, rich become richer. However, in this plot, either end doesn’t mean a good thing, because you either don’t qualify to get a credit (thus 0 BankcardUtilization), or you already maxed out your credit.
ProsperRating is the most important variable affecting LenderYield, the IQR of lender yield of different categories are never overlap, and even with some outliers, the mean lender yield of each category still close to its median. So for most of the time, a borrower at better category (like AA) will get a better (lower) interest rate than those at inferior category (like A). Two numbers can further support the plot, their correlation is about -0.95, and if we run a linear regression model using formula = LenderYield ~ ProsperRating, we can get about 0.91 of r-square.
If we look at the math formula of how to calculate interest rate, under the same loan amount and the same term, higher interest rate gets higher monthly payment, you would think that monthly payment should have positive correlation with interest rate (LenderYield). However, if we look at the big picture, the LenderYield actually has negative correlation with MonthlyLoanPayment (-0.33). In this case, higher ProsperRating gets lower interest rate and higher LoanOriginalAmount, which equal to MonthlyLoanPayment, that’s why LenderYield is negatively correlated with MonthlyLoanPayment.
Prosper loan dataset is a big dataset, 81 features, each represents something you might or might not ignore, depends on what you try to understand. Initially, I want to figure out what features affecting lender yield, because I think this is an important number investor want to know, this number should include everything such as risk premium. The process of investigation spent more time than I was expected. One of the reasons is that I tried to build a linear regression model using fewer features to achieve higher r-squared. Some variables seem important intuitively, however, once included in the linear regression model, it only improves r-squared by 0.012 (Listing Category). So I spend most times reading and thinking, it’s rewarding though, not only have I gained the knowledge of how Prosper functioning, but also I have learned a lot about how to rate one person’s credit.
During the process of investigation on the dataset and prosper website, I figured out the interest rate is decided by Prosper rating. Most borrowers in the same rating will get a narrow range of interest rate which does not overlap with other ratings.
After finding out the fact, my question soon to become: what features deciding Prosper rating? Prosper doesn’t reveal how they calculate the rating of borrowers; it’ll be fascinating if we can approximate the result using the data on hand. I think that this is the fun part of data analyze, you try to discover the insight from the data, predict one of the features.
And through the process, I also learned some of the facts like most borrowers only borrow money that less than three times of their monthly income. I think this is another business insight worth to figure out: the reason people come to Prosper. Combined with the fact that 63% of the purpose of the borrowing is debt consolidation, and the interest rate of Prosper is slightly lower than credit card rate. I can think of one reason that people come to Prosper is to get a lower interest rate for their credit card debt. This may provide some room for arbitrage if we know our audiences well.